To make matters worse, there are often insufficient people within the tendering organisation who know enough about what they are buying, which is how you end up with unreliable or unsuitable trains, as with the Fyra fiasco.

There are also some worrying trends in the way equipment is being procured. One example is the framework contract, where a railway goes out to tender for a fleet of trains as part of a much larger contract for trains to be acquired in additional batches over several years. This differs from the more common practice of including options for additional equipment under similar terms but with an expiry date on exercising the options.

A framework contract is attractive because the operator should be able to secure a better unit price due to economies of scale, while the supplier has continuity of production for several years.

Unfortunately, there is often no certainty that the buyer will take up its options for additional trains, even beyond the initial batch. But where should the supplier pitch its bid in terms of the price? Should it play it safe and assume that only the initial batch of trains will be procured and risk losing a potentially lucrative contract, or should it be brave and assume that a large number of trains will be procured? This quandary affects not just the main contractor but also filters down through the large number of sub-contractors in the supply chain.

There are two dangers here: the buyer ends up paying more per unit than necessary because suppliers were not willing to pitch their bids low enough, or the winner has bid on the assumption that the entire fleet will be ordered, only to find that the customer walks away from the deal after the first batch of trains has been supplied. Here, the customer gets some very cheap trains, and the supplier takes a financial hit. While this is a risk of doing business, there will be negative consequences from this type of buyer behaviour. Manufacturers need to remain profitable to stay in business, and operators need a reasonable number of suppliers for competition to work effectively.

Another trend is the move away from staged payments to payment in full only when the order has been completed to the customer's satisfaction. I suspect the motive for this change is the inordinate time it takes in Europe to certify trains for operation. Perhaps some operators believe this will encourage manufacturers to build trains which will sail through the labyrinthine certification process. Most of the delays in obtaining approvals are not due to large numbers of faults being discovered but because the certification process is no longer fit for purpose.

Staging payments has benefits on both sides. The railway can tie in the payments to key mileposts during manufacture and withhold a payment if it is unhappy with the build quality, for example. The system also acts as an incentive to the manufacturer to keep the contract on schedule.

Insisting on payment in full on completion of delivery has consequences. The railway could end up paying higher prices if more risk is transferred to cover the increased cost of funding the deal. "We are not bankers and we have no desire to be," one frustrated leading supplier told me recently.

Unfortunately, this is exactly what is happening with two of Britain's largest train procurement projects, where the Department for Transport (DfT) has insisted that Hitachi and Siemens and their partners fund the manufacture of the trains and then supply and maintain them over a fixed term. This is a very expensive way to purchase trains as the supplier has to bear all the risk for funding, developing, manufacturing and maintaining the vehicles. These contracts were so complicated to negotiate that it took nearly three-and-a-half years to conclude the Hitachi deal, partly because the DfT kept changing its mind about what type of train it wanted, and two years for the Siemens contract. This in itself increases the cost of procurement.

The high cost of funding, developing, producing and maintaining the first batch of Hitachi trains is evident from an average cost per vehicle of £7.7m compared with £4.4m for the additional cars ordered last month. Even though the first order is for a mixture of electric and bi-mode trains whereas the new order is only for electric trains, this is still a huge difference in cost.

Failure to understand where the boundaries lie in the relationship between buyer and supplier will have unforeseen and potentially disastrous results, which is why it is vital for those responsible for procurement to be informed buyers and to consider the consequences of their actions.